Diagnostic problem-solving using fuzzy abductive inference with degrees of manifestations

摘要

Abduction is one of the methods of inference for creative thinking support Systems or diagnosis support systems. Peng and Reggia proposed abductive inference based on associations between causes and effects for fault diagnosis. Their method hadthe advantages in the calculation cost for of explanations. They discussed the probabilistic relations between the disorders and the manifestations for the evaluation of the explanations. However, it is required for the real world fault/medical diagnosisto treat the degree of manifestation. There are cases where that the probability of the calculated explanation is low but its belief which is calculated based on the degrees of manifestation is high. In this paper we propose a diagnostic problem-solvingmethod using fuzzy abductive inference which can treat the degree of manifestation. This method can infer the combinations of disorders with their degrees of belief using the degrees of manifestations. This method provides the flexible description of thetarget system using the degrees of disorders by linguistic labels and the necessary cause-and-effect relationships.